When running an execution and it fails in a Tabular, isn't it possible to continue executing the other tables (which are of course not a source for the failing Tabular). Now in our process the execution stops, but the tabular is just a last step for a few tables, not for all other tables.
Solved
Is is possible to 'Continue on error' when an execution is failing?
Best answer by Thomas Lind
Hi
Here is how I do the setup explained by Rory.
The Prod and Dev packages contain the sub packages that is marked in red.
Method 1: Run on success.
The Prod/Dev package is set up like this.
All it does is start the BU package.
The BU package is set up like so.
It just runs the Business Units, but does so in 4 threads compared to 1 for the start package. This also contains failure handling, so it can be handled in this instead of affecting the other parts. The last thing it does is to run the DSA package if it is successful.
This is the same for all packages until the Semantic one is hit.
Method 2: Sub packages is executed by main package.
The Dev package is set up like so.
The difference here is that it has all the other packages added in the Include Steps area. Because it is set to run in one thread without Merge steps and Managed Execution set to disabled, it will run it one step at a time from top to bottom.
Each individual package will be set up like so, which is very similar to the other method, only it does not use the Run Package on success.
The benefit of this method compared to the other, is that you can run one of the sub packages on its own without making it run the other steps on success.
In essence the results are very similar and each section runs on its own.
+7Hi,
in the 20.10.x release you would need to split the Tabular executions to a separate package that runs after parts of the project you want to have more ensurance of execution. You can do this with the option to run a specific package after another package. Essentially, any error will cause the overarching package to fail in nested packages.
In the 6xxx release, you can use Orchestration to build any type of reload sequence you want. There you can specifiy whether or not errors should propagate up the chain.
1 person likes this
+5Hi
Here is how I do the setup explained by Rory.
The Prod and Dev packages contain the sub packages that is marked in red.
Method 1: Run on success.
The Prod/Dev package is set up like this.
All it does is start the BU package.
The BU package is set up like so.
It just runs the Business Units, but does so in 4 threads compared to 1 for the start package. This also contains failure handling, so it can be handled in this instead of affecting the other parts. The last thing it does is to run the DSA package if it is successful.
This is the same for all packages until the Semantic one is hit.
Method 2: Sub packages is executed by main package.
The Dev package is set up like so.
The difference here is that it has all the other packages added in the Include Steps area. Because it is set to run in one thread without Merge steps and Managed Execution set to disabled, it will run it one step at a time from top to bottom.
Each individual package will be set up like so, which is very similar to the other method, only it does not use the Run Package on success.
The benefit of this method compared to the other, is that you can run one of the sub packages on its own without making it run the other steps on success.
In essence the results are very similar and each section runs on its own.
+5Hi
Reply
Most helpful members this week
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.
